커스텀 임베딩(미리보기). 자체 임베딩을 만든 경우 Vertex AI Search에서 생성한 임베딩 대신 이를 사용하여 추가 컨텍스트로 검색을 보강할 수 있습니다. 이 기능은 정형 데이터가 포함된 데이터 스토어 또는 메타데이터가 있는 비정형 데이터에 사용할 수 있습니다. 커스텀 임베딩 사용을 참고하세요.
웹사이트에 검색 위젯을 삽입합니다. Vertex AI Search는 웹사이트 또는 웹 애플리케이션에 복사할 수 있는 코드를 제공합니다. 이렇게 하면 검색 위젯이 배포됩니다. 콘솔에서 검색 결과를 미리 볼 수 있습니다.
웹페이지에 검색 위젯 추가를 참고하세요.
서버 또는 애플리케이션에 검색 API 호출을 통합합니다. 검색 결과가 표시되는 방식을 완전히 제어하려면 API 호출을 서버 또는 애플리케이션에 통합하면 됩니다. API 호출에 관한 자세한 내용은 검색 결과 가져오기를 참고하세요. 클라이언트 라이브러리 리소스는 AI 애플리케이션 클라이언트 라이브러리를 참고하세요.
[[["이해하기 쉬움","easyToUnderstand","thumb-up"],["문제가 해결됨","solvedMyProblem","thumb-up"],["기타","otherUp","thumb-up"]],[["이해하기 어려움","hardToUnderstand","thumb-down"],["잘못된 정보 또는 샘플 코드","incorrectInformationOrSampleCode","thumb-down"],["필요한 정보/샘플이 없음","missingTheInformationSamplesINeed","thumb-down"],["번역 문제","translationIssue","thumb-down"],["기타","otherDown","thumb-down"]],["최종 업데이트: 2025-09-03(UTC)"],[[["\u003cp\u003eThis page outlines the steps to create a generic search application using Vertex AI Agent Builder, guiding users from setting up a Google Cloud project to deploying their search app.\u003c/p\u003e\n"],["\u003cp\u003eThe process includes preparing and importing data, whether unstructured (like PDFs) or structured (like BigQuery tables), and then creating a data store to house this data for searching.\u003c/p\u003e\n"],["\u003cp\u003eUsers can configure search settings, which vary based on whether they deploy a search widget or integrate search API calls, allowing customization of results, autocomplete, and more.\u003c/p\u003e\n"],["\u003cp\u003eThe platform offers options to preview the search results and refine configurations before deploying the app either by embedding a widget or through API integration into a server or application.\u003c/p\u003e\n"],["\u003cp\u003eOnce deployed, the guide provides information on maintaining the app, including how to refresh data to ensure search results are current and accurate.\u003c/p\u003e\n"]]],[],null,["# Custom search checklist\n\nThis page provides a checklist of the steps required to create a custom search app. For a checklist to create a website search app, see [Website search checklist](/generative-ai-app-builder/docs/website-search-checklist).\n\n\u003cbr /\u003e\n\nIf you're new to AI Applications, consider following the [Get started\nwith Vertex AI Search](/generative-ai-app-builder/docs/try-enterprise-search) tutorial to\ncreate a sample app.\n\n\n### [Set up a Google Cloud\nproject](#set-up-project)\n\nSet up a Google Cloud project, turn on AI Applications, and set up access\ncontrol for your project. You can use an existing Google Cloud project if you\nhave one already.\n\n### Actions\n\n1. Review [Before you begin](/generative-ai-app-builder/docs/before-you-begin) and confirm that you have completed the steps. \n\n### [Prepare your data](#prepare-data)\n\nDetermine what kind of data you'll make available in your search app and prepare\nit for importing to Vertex AI Search.\n\nYou can base your data store on the following types of data:\n\n- **Unstructured data.** Documents (such as PDFs, HTML files, and TXT files) that are stored in Cloud Storage. Optionally, you can provide metadata in a JSON file or in a BigQuery table.\n- **Structured data.** Data provided with a specific schema. For example, you can provide data in a BigQuery table, as JSON files in Cloud Storage, or from third-party connectors such as Jira.\n\n### Actions\n\n1. Review the information about supported data and the relationship between apps\n and data stores in [About apps and data\n stores](/generative-ai-app-builder/docs/create-datastore-ingest).\n\n2. Prepare your data according to the requirements in [Prepare data for\n ingestion](/generative-ai-app-builder/docs/prepare-data).\n\n3. If you need to set up access control to limit the data that users can view in\n your search app's results, review the prerequisites and follow the\n instructions for your identity provider type and data source in [Use data\n source access control](/generative-ai-app-builder/docs/data-source-access-control).\n\n### [Import your data](#import-data)\n\nCreate a data store and then import your data into it, or set up a\nthird-party connector as your synced data source.\n\nHow you import your data depends on where you're importing it from. For\nexample, if your data is in Cloud Storage, you can import it using the\nconsole or the API by providing the bucket location of your data.\n\n### Actions\n\n1. Follow the instructions for your data source in [Create a search data\nstore](/generative-ai-app-builder/docs/create-data-store-es). \n\n### [Create your app](#create-app)\n\nCreate your custom search app and connect it to your new data store.\n\n### Actions\n\n1. [Create a search app](/generative-ai-app-builder/docs/create-engine-es). \n\n### [Configure your search\nsettings](#configure-settings)\n\nVertex AI Search provides many configuration options. Some options\ndepend on whether you plan to deploy a search widget.\n\n### Actions\n\nDepending on your use case and whether you plan to deploy the out-of-the-box\nsearch widget or integrate search API calls into your own code,\nVertex AI Search provides several options for configuration.\n\n1. You can embed a search widget into your website. The widget automatically\n provides a search bar and expandable search interface. If you plan to deploy\n the search widget, configure the following:\n\n - **Search widget results.** See [Configure results for the\n search widget](/generative-ai-app-builder/docs/configure-widget-attributes).\n\n - **Search widget facets (Preview).** See [Configure facets\n for the search widget](/generative-ai-app-builder/docs/configure-widget-facets).\n\n2. If you plan to integrate search API calls into your server or application\n instead of using the widget, you can configure your search settings using\n the following options:\n\n - **Field settings.** For structured data or for unstructured data with\n metadata, update field settings to refine how Vertex AI Search\n uses metadata for search. See [Configure fields for\n search](/generative-ai-app-builder/docs/configure-search-attributes).\n\n - **Autocomplete.** Depending on your data, set up autocomplete suggestions\n based on document content, fields, search history, or user events. See\n [Configure autocomplete](/generative-ai-app-builder/docs/configure-autocomplete).\n\n - **Serving controls.** Control when search results\n are boosted, buried, filtered, or redirected, or whether certain queries\n are associated with other queries. See [Configure serving\n controls](/generative-ai-app-builder/docs/configure-serving-controls).\n\n - **Search tuning (Preview).** Tune the search model with your own training\n data. See [Improve search results with search\n tuning](/generative-ai-app-builder/docs/tune-search).\n\n - **Custom embeddings (Preview).** If you've created your own embeddings,\n you might prefer to use them instead of those generated by\n Vertex AI Search to enrich your searches with additional\n context. This feature is available for data stores with structured data or\n unstructured data with metadata. See [Use custom\n embeddings](/generative-ai-app-builder/docs/bring-embeddings).\n\n### [Configure your search\nresults](#configure-results)\n\nIf you plan to deploy your app by integrating search API calls into your own\ncode, Vertex AI Search provides additional options for configuring how\nyour search results are returned.\n\n### Actions\n\n1. Configure your search results with the following options:\n\n - [Filter custom search for structured or\n unstructured data](/generative-ai-app-builder/docs/filter-search-metadata).\n - **Get snippets, extractive answers, or extractive segments.** See [Get\nsnippets and extracted content](/generative-ai-app-builder/docs/snippets). \n\n### [Preview search\nresults](#preview-results)\n\nYou can preview your search results to check if your app configurations are\nworking as expected.\n\n### Actions\n\n1. To preview your search results, use the console or the API.\n\n - **Console.** Use the AI Applications console **Preview** page to preview\n how search widget configurations affect your results. See the **Console**\n instructions in [Get search\n results](/generative-ai-app-builder/docs/preview-search-results).\n\n - **API** . If you're integrating API calls into your application, make API\n calls to preview your search configurations. See the **REST**\n instructions in [Get search\n results](/generative-ai-app-builder/docs/preview-search-results).\n\n### [Deploy your search\napp](#preview-results)\n\nWhen you're happy with the preview version of your search app, share it with\nyour users by deploying it to your website.\n\n### Actions\n\n1. You can deploy your search app in either of the following ways:\n\n - **Embed the search widget into your website.** Vertex AI Search\n provides code that you can copy into your website or web application. This\n deploys the search widget. You can preview search results in the console.\n See [Add the search widget to a webpage](/generative-ai-app-builder/docs/add-widget).\n\n - **Integrate search API calls into your server or application.** For full\n control over how your search results are displayed, you can integrate API\n calls into your server or applications. For more information about making\n API calls, see [Get search results](/generative-ai-app-builder/docs/preview-search-results). For client library resources, see [AI Applications client\n libraries](/generative-ai-app-builder/docs/libraries).\n\n### [Record and import user events](#user-events)\n\nTo get personalized search and browse results, you can update the user events\nin your search app. For more information see\n[About user events](/generative-ai-app-builder/docs/user-events).\n\n### Actions\n\n1. [Import historical user events](/generative-ai-app-builder/docs/import-user-events).\n2. [Record real-time user events](/generative-ai-app-builder/docs/record-user-events). \n\n### [Maintain your app](#maintain-data)\n\nYou can maintain your app to ensure that latest and necessary data is available\nin your data store.\n\n### Actions\n\n1. [Refresh your data](/generative-ai-app-builder/docs/refresh-data)."]]
